Model save

README.md

@@ -2,13 +2,11 @@
 license: llama2
 library_name: peft
 tags:
-- alignment-handbook
 - trl
 - sft
+- alignment-handbook
 - generated_from_trainer
 base_model: meta-llama/Llama-2-7b-hf
-datasets:
-- HuggingFaceH4/ultrachat_200k
 model-index:
 - name: llama2-20p-POE
   results: []
@@ -19,9 +17,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 # llama2-20p-POE
 
-This model is a fine-tuned version of [meta-llama/Llama-2-7b-hf](https://huggingface.co/meta-llama/Llama-2-7b-hf) on the HuggingFaceH4/ultrachat_200k dataset.
-It achieves the following results on the evaluation set:
-- Loss: nan
+This model is a fine-tuned version of [meta-llama/Llama-2-7b-hf](https://huggingface.co/meta-llama/Llama-2-7b-hf) on the None dataset.
 
 ## Model description
 
@@ -56,9 +52,6 @@ The following hyperparameters were used during training:
 
 ### Training results
 
-| Training Loss | Epoch | Step | Validation Loss |
-|:-------------:|:-----:|:----:|:---------------:|
-| 0.7327        | 1.0   | 1039 | nan             |
 
 
 ### Framework versions

adapter_config.json

@@ -19,12 +19,12 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "k_proj",
-    "o_proj",
     "up_proj",
-    "gate_proj",
-    "down_proj",
+    "k_proj",
     "q_proj",
+    "down_proj",
+    "gate_proj",
+    "o_proj",
     "v_proj"
   ],
   "task_type": "CAUSAL_LM"

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:af8b59a0f339223195ffda22e2cc190e5b99e802340991131a98cce499515eaf
+oid sha256:d370cf8d66f04acedfbae1d4d8d05426bcf20615ba409e1ded0bd718bca76c0f
 size 60089544

all_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 1.0,
-    "train_loss": 0.7603742487735766,
-    "train_runtime": 32307.7024,
-    "train_samples": 33257,
-    "train_samples_per_second": 1.029,
-    "train_steps_per_second": 0.032
+    "epoch": 0.96,
+    "train_loss": 0.0,
+    "train_runtime": 0.0161,
+    "train_samples": 27020,
+    "train_samples_per_second": 1676083.975,
+    "train_steps_per_second": 52354.363
 }

runs/May03_15-11-58_ip-172-31-69-60.ec2.internal/events.out.tfevents.1714749305.ip-172-31-69-60.ec2.internal.10862.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b357e01c92dc4090b676eaa1e8347503d8b72698ed29bff76b0104920d488352
+size 5066

tokenizer.json

@@ -1,6 +1,11 @@
 {
   "version": "1.0",
-  "truncation": null,
+  "truncation": {
+    "direction": "Right",
+    "max_length": 2048,
+    "strategy": "LongestFirst",
+    "stride": 0
+  },
   "padding": null,
   "added_tokens": [
     {

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 1.0,
-    "train_loss": 0.7603742487735766,
-    "train_runtime": 32307.7024,
-    "train_samples": 33257,
-    "train_samples_per_second": 1.029,
-    "train_steps_per_second": 0.032
+    "epoch": 0.96,
+    "train_loss": 0.0,
+    "train_runtime": 0.0161,
+    "train_samples": 27020,
+    "train_samples_per_second": 1676083.975,
+    "train_steps_per_second": 52354.363
 }

trainer_state.json

@@ -1,9 +1,9 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 0.9995189995189995,
+  "epoch": 0.962000962000962,
   "eval_steps": 500,
-  "global_step": 1039,
+  "global_step": 1000,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
@@ -1416,78 +1416,21 @@
       "step": 1000
     },
     {
-      "epoch": 0.97,
-      "grad_norm": 0.10404789525698893,
-      "learning_rate": 6.518264651449779e-07,
-      "loss": 0.7771,
-      "step": 1005
-    },
-    {
-      "epoch": 0.97,
-      "grad_norm": 0.09905350029610553,
-      "learning_rate": 4.743499715179067e-07,
-      "loss": 0.7495,
-      "step": 1010
-    },
-    {
-      "epoch": 0.98,
-      "grad_norm": 0.08940324837180028,
-      "learning_rate": 3.249628053142884e-07,
-      "loss": 0.7587,
-      "step": 1015
-    },
-    {
-      "epoch": 0.98,
-      "grad_norm": 0.10668411274319571,
-      "learning_rate": 2.0370712837906037e-07,
-      "loss": 0.7657,
-      "step": 1020
-    },
-    {
-      "epoch": 0.99,
-      "grad_norm": 0.0933938060668371,
-      "learning_rate": 1.1061716294951118e-07,
-      "loss": 0.7493,
-      "step": 1025
-    },
-    {
-      "epoch": 0.99,
-      "grad_norm": 0.09579015452060988,
-      "learning_rate": 4.5719181996650705e-08,
-      "loss": 0.7677,
-      "step": 1030
-    },
-    {
-      "epoch": 1.0,
-      "grad_norm": 0.1038186774195005,
-      "learning_rate": 9.031501810174981e-09,
-      "loss": 0.7327,
-      "step": 1035
-    },
-    {
-      "epoch": 1.0,
-      "eval_loss": NaN,
-      "eval_runtime": 3002.6712,
-      "eval_samples_per_second": 0.769,
-      "eval_steps_per_second": 0.192,
-      "step": 1039
-    },
-    {
-      "epoch": 1.0,
-      "step": 1039,
-      "total_flos": 3.479777588753203e+16,
-      "train_loss": 0.7603742487735766,
-      "train_runtime": 32307.7024,
-      "train_samples_per_second": 1.029,
-      "train_steps_per_second": 0.032
+      "epoch": 0.96,
+      "step": 1000,
+      "total_flos": 3.3476112847732736e+16,
+      "train_loss": 0.0,
+      "train_runtime": 0.0161,
+      "train_samples_per_second": 1676083.975,
+      "train_steps_per_second": 52354.363
     }
   ],
   "logging_steps": 5,
-  "max_steps": 1039,
+  "max_steps": 844,
   "num_input_tokens_seen": 0,
   "num_train_epochs": 1,
   "save_steps": 100,
-  "total_flos": 3.479777588753203e+16,
+  "total_flos": 3.3476112847732736e+16,
   "train_batch_size": 4,
   "trial_name": null,
   "trial_params": null

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2fa7a10ac6136e7849c5c631b07d76c63055ef105757f7af1a89431085e9e7bb
+oid sha256:abedc56d858f39be9d29990c2dd98aa13e928d1420e5e2f5c14c5d669d84e4a4
 size 6072